Automation might someday allow for the inexpensive creation of highly contextualized and effective laws. If that ever comes to pass, however, it will not be on a blank slate. Proponents will face the question of how to computerize bedrock aspects of our existing law, some of which are legal standards—norms that use evaluative, even moral, criteria. Conventional wisdom says that standards are difficult to translate into computer code because they do not present clear operational mechanisms to follow. If that wisdom holds, one could reasonably doubt that legal automation will ever get off the ground.
Conventional wisdom, however, fails to account for the plasticity of law. The murkiness of standards makes them a fertile ground for the growth of competing interpretations of their legal meaning. Some of those interpretations might be more rule-like than others and, therefore, easier to automate. Proponents of automation will likely be drawn to those rule-like interpretations, so long as they are compatible enough with existing law.
This complex dynamic between computer-friendliness and the changing content of law makes it troublesome for legislators to identify the variable and fixed costs of automation. This Article aims to shed light on this relationship by focusing our attention on a quintessential legal standard at the center of our legal system—the “reasonably prudent person test.”
Here, I explain how automation proponents might be tempted by fringe, formulaic interpretations of the test, such as Averageness, because they bring comparatively low innovation costs. With time, however, technological advancement will likely drive down innovation costs, and mainstream interpretations, like Conventionalism, could find favor again. Regardless of the interpretation that proponents favor, though, an unavoidable cost looms: by replacing the jurors who apply the test with a machine, they will eliminate a long-valued avenue for participatory and deliberative democracy.